Mozer, M. C.: Neural network music composition by prediction: Exploring the benefits of psychoacoustic constraints and multi-scale processing. Connection Science 6 (1994) 247280. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Natural Language Processing - Hirzel, Soukup (2000) (Correct)
....as far as we can see, Markov chains have always been used in greedy algorithms, while we are going to use them with dynamic programming. Since Laske [7] many researchers have used formal grammars for music generation. Neural networks have been trained for this problem, as described in Mozer [10]. Others used constraint based and rule based systems. We will present in greater detail some of these methods in a section 2. The organization of the remainder of this write up is the following: in the next section we present a survey of methods that have been used in algorithmic composition. ....
....towards the musical text. It would be better, therefore, to call them derivational. 2. 2 Neural Networks Making the argument that the connectionist approach o ers algorithms that discover relevant structure, noise resistance and better generalization through distributed representations, Mozer [10] pursues a connectionist approach to algorithmic composition. Mozer [10] describes a connectionist network that uses a set of melodies written in a given style to compose new melodies in that style. The network used in Mozer s work is an extension of a traditional algorithmic composition ....
[Article contains additional citation context not shown here]
Michael Mozer. Neural network music composition by prediction: Exploring the benets of psychoacoustic constraints and multiscale processing. Connection Science, 1994.
Creating Melodies with Evolving Recurrent Neural Networks - Chen, Miikkulainen (2001) (Correct)
....of rules about the cell structures characteristic to Bartok. We believe that the results from this system will sound, to an extent, like Bartok s music. Artificial composition has been an active research area for a long time. Techniques involving perception [8] emotion [13] and psycho acoustics [11] have been proposed. These approaches presumed a certain state of mind and tried to approximate the stream of thought while a composer is writing his or her music. While the results were promising, they often lacked flexibility in generating variations in the melody, and they drew little support ....
....the stream of thought while a composer is writing his or her music. While the results were promising, they often lacked flexibility in generating variations in the melody, and they drew little support from music theory. Other approaches were more scientific in that they used a statistical model [11] or a knowledge base [1] to estimate or predict each step in the composition. These models worked well in emulating specific style of composition, but they were not able to generate variation not predefined in the database. Recently, Todd and Loy [15] and Desain [5] used a connectionist approach ....
Mozer, M. C. (1994). Neural Network Music Composition by Prediction: Exploring the Benefits of Psychoacoustic Constraints and Multi-scale Processing. Connection Science 6:247--80.
Computing Auditory Perception - Purwins, Blankertz, Obermayer (2000) (1 citation) (Correct)
....given a melody. 4. COGNITION AND PERCEPTION IN COMPOSITION 4.1. Automated Composition The use of neuro mimetic artificial neural nets for music composition is only partially successful. Melody generation with the backpropagation through time algorithm with perceptually relevant preprocessing (Mozer 1994), as well as using stochastic Boltzmann machines for choral harmonization (Bellgard and Tsang 1994) does not yield musically pleasant results. They do not reach the quality of compositions generated by the elaborated rule based system in (Cope 1996) Bach choral harmonization with a simple ....
Mozer, M. C. 1994. Neural network music composition by prediction: exploring the benefits of psychoacoustic constraints and multi-scale processing. Connection Science 6(2 & 3): 247-80.
Handling Time-Warped Sequences with Neural Networks - Ulbricht (1996) (7 citations) (Correct)
....new input are combined to form the state of the network, the first layer of the multi layer perceptron is called the state layer, and the network is called the input state network. The network is depicted in Fig. 4. It is close to the network models described in [Mozer and Soukup, 1991] and [Mozer, 1994], but it differs in several respects. One important difference is that, in these networks, the weights of the connections leading to the state layer are trainable. It is also different from the network presented in [Mozer, 1992] where the units in the hidden layer are equipped with temporal ....
M.C. Mozer. Neural Network Music Composition by Prediction: Exploring the Benefits of Psychoacoustic Constraints and Multiscale Processing. Connection Science, 1994.
A Supervised Learning Approach to Musical Style Recognition - Buzzanca (2002) (Correct)
No context found.
Mozer, M. C.: Neural network music composition by prediction: Exploring the benefits of psychoacoustic constraints and multi-scale processing. Connection Science 6 (1994) 247280.
GP-Music: An Interactive Genetic Programming System for Music .. - Johanson, Poli (1998) (1 citation) (Correct)
No context found.
Mozer, M.C., et al., "Neural network music composition by prediction: exploring the benefits of psychoacoustic constraints and multi-scale processing," Connection Science,. Vol.6, No.2-3, pp. 247-80, 1994.
Learning Harmonic Progression Using Markov Models - Clement (1998) (Correct)
No context found.
M.C. Mozer. Neural Network Music Composition by Prediction: Exploring the Benefits of Psychoacoustic Constraints and Multi-Scale Processing. Connection-Science. vol. 6, no. 2-3; pp.247-280, 1994.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC